Cell Transplantation, Vol. 22, pp. 1–14, 2013 Printed in the USA. All rights reserved. Copyright 2013 Cognizant Comm. Corp.
0963-6897/13 $90.00 + .00 DOI: http://dx.doi.org/10.3727/096368912X653282 E-ISSN 1555-3892 www.cognizantcommunication.com
Review Cardiac Stem Cell Therapy: Stemness or Commitment? Ashish Mehta and Winston Shim Research and Development Unit, National Heart Centre Singapore, Singapore
Cardiac stem cell therapy to promote engraftment of de novo beating cardiac muscle cells in cardiomyopathies could potentially improve clinical outcomes for many patients with congestive heart failure. Clinical trials carried out over the last decade for cardiac regeneration have revealed inadequacy of current approaches in cell therapy. Chief among them is the choice of stem cells to achieve the desired outcomes. Initial enthusiasm of adult bone marrow stems cells for myocyte regeneration has largely been relegated to paracrine-driven, donor cell-independent, endogenous cardiac repair. However, true functional restoration in heart failure is likely to require considerable myocyte replacement. In order to match stem cell application to various clinical scenarios, we review the necessity to preprime stem cells towards cardiac fate before myocardial transplantation and if these differentiated stem cells could confer added advantage over current choice of undifferentiated stem cells. We explore differentiation ability of various stem cells to cardiac progenitors/cardiomyocytes and compare their applicability in providing targeted recovery in light of current clinical challenges of cell therapy. Key words: Stem cells; Cardiomyocytes; Differentiation; Myocardial infarction; Regenerative medicine
INTRODUCTION During mammalian development, embryonic heart expands through proliferation of its constituent cell types of which 60% are cardiac fibroblasts, 30% force-generating heart muscle cells (cardiomyocytes), and the remainders, vascular endothelial and smooth muscle cells (31). Proliferative capacity of those cardiomyocytes rapidly declines and becomes terminally differentiated shortly after birth (2). Although reports suggest the presence of progenitor populations in the heart (67), their ability to support basal turnover of cardiomyocytes is debatable (56). Critically, when large numbers of cardiomyocytes are lost as in myocardial infarction, intrinsic cardiac regeneration is invariably poor. Current effective therapy to prevent total congestive heart failure demands orthotropic transplantation, but shortage of donor organs, high costs, and complications of immunosuppressants continue to plague the rehabilitation process. In recent years, application of cell-based therapy has gained considerable interest as an alternate means to “remuscularize” the injured heart and to overcome limitation of organ shortage. Early preclinical studies provide evidence of fetal, and neonatal cardiomyocytes proliferate
and survive when injected in the heart (59,76,96). More importantly, these cells repopulated scars, buttressing wall thickness and eventually improved left ventricular function (76,126) with signs of electromechanical integration (92,134). These initial results suggested that cardiomyocyte transplantation may be a promising avenue to replace fibrotic scar. However, use of fetal and neonatal cardiomyocytes poses significant clinical hurdle due to their limited supply and ethical concerns. These have fueled the search for alternative cardiogenic source for myocardial regeneration (Table 1). The first experimental attempts to restore cardiac function in an injured myocardium utilized multipotent cells from skeletal muscle and bone marrow. Although not bona fide heart stem cells, accessibility, scalability, and autologous source, make them viable candidates (39). Unfortunately, experiments in rodents demonstrated that myoblasts transplanted to infarcted hearts survived but failed to convert to cardiomyocytes (89) and demonstrated no electromechanical coupling (93). However, some groups (36,112) demonstrated positive effects on heart function through poorly understood mechanisms. Furthermore, randomized, double-blind,
Received October 8, 2011; final acceptance February 18, 2012. Online prepub date: August 27, 2012. Address correspondence to Dr. Winston Shim, National Heart Centre Singapore, 17, Third Hospital Avenue, Mistri Wing, Singapore 168752. Tel: +65 64350752; Fax: +65 62263972; E-mail: [email protected]
2 mehta and shim
Table 1. Various Types of Stem Cells With Applications in Cardiac Therapy Type
Isl-1+ Sca-1+ CSC
Heart Heart Heart
Differentiation to CM
CD90 , CD45 , CD117 , CD29 CD44+, CD90+, CD106+, CD71+, CD45low, CD133+, CXCR4+ CD34+, CD31+, CD146+, vWF+, CD45low, VE-CAD+, CXCR4+ Sca1+, CD34+, Nkx2.5+, GATA4+, CD45–, Oct4+, CXCR4+ Isl1+, Nkx2.5+, Oct4–, GATA4+ +
Isl1+, Nkx2.5+, GATA4+, α-MHC+ Sca1+, Nkx2.5+, cTnI+ MDR+, KDR+, CD31+, cTnI+, Nkx2.5– cKit+, Isl1+
Treatment with 5Aza-dC Treatment with 5AzadC and coculture with CM Treatment with 5Aza-dC
Spontaneous and GF-induced differentiation Spontaneous differentiation Treatment with 5Aza-dC Spontaneous differentiation and low frequency electromagnetic waves Treatment with 5Aza-dC
(55) (81) (72) (109)
HSC, hematopoietic stem cell; MSC, mesenchymal stem cell; EPC, endothelial progenitor cell; VSEL, very small embryonic-like stem cell; ESC, embryonic stem cell; CSC, cardiac stem cell; BM, bone marrow; 5Aza-dC, 5-aza-2¢deoxycytidine; GF, growth factor; CM, cardiomyocyte.
placebo-controlled clinical trial failed to show improvement in several key parameters of cardiac function after a follow-up period of 6 months (71). Bone marrow-derived stem cells, on the other hand, have been demonstrated to differentiate into various cell types including cardiomyocytes. Some groups (10,30,46) suggested bone marrow cells transdifferentiated to cardiomyocytes, whereas others (3,7,79,80,113) attributed such observations to cell fusion between host and transplanted cells. Furthermore, clinical trials often with conflicting results accentuated the controversy (98). Positive effects from bone marrow stem cell transplant have been relegated to poorly understood paracrine effect, possibly by its direct effect on the myocardium or by contributing to increased vascularization (39). Although the field has witnessed considerable controversies over the past decades, cell-based cardiac therapy nevertheless remains a mainstream experimental concept that can revolutionize management of heart failure. In this review, we expound probable stem cell populations for cardiac repair, and we discuss the concept of milieudirected differentiation versus predifferentiation of bone marrow-derived stem cells and human pluripotent stem cells towards muscular lineages in exacting posttransplant benefits in the infarcted myocardium. Exten sive literature on stem cells for cardiac repair has been reviewed elsewhere; we selectively discussed literature on the subject of stem cell cardiomyogenesis in relation to functional benefits. CARDIOMYOGENIC DIFFERENTIATION OF STEM CELLS Traditionally, stem cells consist of two broad categories of adult stem cells and embryonic stem cells (ESCs). While adult stem cells are derived from postnatal somatic
tissues and are generally considered to be multipotent, their embryonic counterparts are derived from the inner cell mass of blastocyst-stage embryos and are pluripotent (114). Recent advancements in stem cell biology has enabled reprogramming of differentiated somatic cell types into a pluripotent state similar to ESCs via induced expression of selected pluripotent genes to yield induced pluripotent stem cells (iPSCs) (108). In the following sections, we discuss cardiogenic ability of those stem cells, and if predifferentiating them to cardiac lineages for exogenous cell-based therapy is warranted. Bone Marrow-Derived Stem Cells and Cardiogenesis Human bone marrow is the major source of adult stem cells. Bone marrow contains a complex assortment of progenitor cells, including hematopoietic stem cells (HSCs), side population (SP) cells, mesenchymal stem cells (MSCs) or stromal cells, and multipotent adult progenitor cells (MAPCs), a subset of MSCs (47). There is considerable evidence to suggest that those stem cell populations harbor putative cardiac transdifferentiation potential (22,65,82,128). Although some of the apparent transdifferentiation events has later been attributed to technical artifacts or previously overlooked phenomena, such as cell fusion and “ectopic” stem cells (3,120). Mesenchymal Stem Cells In the past decade, emphasis of in vitro targeted differentiation towards myocytes has been focused on MSCs, mainly for their ease of isolation and expansion in culture. Makino et al. (65) was among the first to report that immortalized murine MSCs exposed to 5-aza-deoxycytidine (5-aza-dC), an inhibitor of DNA methylation, resulted in appearance of spontaneously beating foci. Furthermore, those cells also expressed hallmark markers of cardiac
DOES DIFFERENTIATION MAKE A DIFFERENCE IN CARDIAC REPAIR?
phenotype. Similarly, Shiota et al. (102), Antonitsis et al. (5), and many other groups (42,86,91) also demonstrated cardiac differentiation of MSCs utilizing 5-aza-dC based on phenotypic observations but without contracting foci. To obviate safety concerns of 5-aza, we induced cardiogenic differentiation from human MSCs using a combination of insulin, dexamethasone, and ascorbic acid to generate cardiomyocyte-like cells (CLCs). They were found to express multiple sarcomeric proteins [cardiac troponin I (cTnI), sarcomeric tropomyosin, and cardiac titin] that are associated with cardiomyocytes. Furthermore, CLCs showed a nascent cardiomyocyte phenotype with cross-striated myofibrils characterized by a-actinin-positive Z bands (99). However, they did not beat spontaneously. Similarly, Schittini et al. (95) utilized conditioned medium from human cardiac explants (HCEs) as a potential source of paracrine factors for differentiating MSCs. Their proteome analysis indicated that HCEs release macromolecules, including cytokines, growth factors, and myocardial and metabolism-related proteins, which trigger differentiation of MSCs to CLCs (95), without contracting foci. Recently, Behfar and colleagues (8) demonstrated that human MSCs could be guided to cardiopoiesis by recombinant cocktail of transforming growth factor-b1 (TGF-b1), bone morphogenetic protein-4 (BMP4), activin A, retinoic acid, insulinlike growth factor-1 (IGF1), fibroblast growth factor-2 (FGF2) and a-thrombin, and interleukin-6 (IL-6). These transformed human cells expressed NK2 transcription factor-realted locus 5 (Nkx-2.5), T-box transcription factor 5 (TBX5), mesoderm posterior 1 (MESP1), and myocyte enhancer factor 2C (MEF2C) and demonstrated synchronized contractions with rhythmic calcium transients in response to external electrical stimuli. More recently, Shinmura et al. (101) reported that treatment of human MSCs with piolitazone, a peroxisome proliferator-activated receptor g (PPARg) activator, significantly enhanced the cardiomyogenic transdifferentiation (BM vs. pBM: 1.9 ± 0.2% n = 47 vs. 39.5 ± 4.7% n = 13, p 30% cardiomyocytes (54). The efficiency of the protocol is not limited to hESCs but works comparatively well with hiPSCs (97,108). Although this protocol generates cardiomyocytes efficiently, its reproducibility across different hESC lines is still unclear (21). The third protocol developed by Keller’s group not only exploits TGF-b family members but also demonstrates the importance of canonical Wnt signaling in cardiogenesis. In Keller’s staged protocol, the combination of activin A and BMP4 induces primitive streak-like population and mesoderm development. Wnt inhibitor, DKK1, is later added to specify cardiac mesoderm and finally vascular endothelial growth factor (VEGF) added to promote expansion and maturation of cardiovascular lineage. This protocol reportedly generates populations consisting of 40–50% cardiomyocytes. Sorting the differentiating cultures for early cardiovascular progenitors further enhanced the efficiency of this protocol. Kinase insert domain receptor/fetal liver kinase (flk1/KDR)low/ ckitneg cells selected on days 5–6 of differentiation when plated as monolayer gave rise to 57% cTnI-positive cells (125). The other population (KDRhigh/ckit+) gave rise to progenitors of hematopoietic and vascular lineages (125). In spite of those defined protocols, majority of lab oratories rely on spontaneous differentiation to generate cardiomyocytes. Spontaneous cardiac differentiation can occur in regular fetal bovine serum containing medium as early as day 10, although efficiency may vary drastically depending on the cell lines used. However, different research groups have developed variations to this basic protocol by adding growth factors or small molecules that further enhance EBs towards cardiac lineage. While some groups provide a serum shock by reducing serum concentrations from 20% to 2% in media (129), other have utilized “spin EBs” (EBs generated from precise cell numbers) to enhance cardiogenic differentiation (14). A summary of some of these methods is presented in Table 2. DIFFERENTIATED OR UNDIFFERENTIATED CELLS FOR CARIDAC REPAIR? In the current section, we focus on the abilities of these cells (undifferentiated vs. differentiated) in attenuating adverse ventricular remodeling. While early studies by Sakai et al. (94) and Chiu (19) showed that bone marrow stem cells performed better than fibroblasts in cardiac cell therapy, one major question remains unanswered is whether cardiac differentiation of bone marrow stem
DOES DIFFERENTIATION MAKE A DIFFERENCE IN CARDIAC REPAIR?
Table 2. Variations in Embryoid Body-Based Differentiation Protocols for Generation of Pluripotent Stem Cell-Derived Cardiomyocytes Group
Burridge et al.
HUES7, BG01, NOTT1, NOTT2
Forced-aggregation of defined numbers of hESCs in V-96 plates, with activin A and bFGF
Xu et al. Zhang et al.
H1, H7, H9, H9.1, H9.2 iPSC lines
Mohr et al.
Gai et al.
H9 and iPSC
Gaur et al. and Mehta et al. Lee et al.
H9 and iPSC lines iPSC lines and H7, HES-3 SNUhES3 and SNUhES4
20% FBS containing media and various differentiating factors 20% FBS containing media for first 10 days followed by 2% FBS media Microwell for EB formation, initial 10 days in 20% FBS followed by 2% FBS containing medium 20% FBS medium for 6 days; followed by 10% FBS medium with various cardiogenic inducers Cardiac differentiation with p38MAPK inhibitor StemPro-34 with a combinations of various growth factors BMP2 addition post attachment of EBs
Differential beating efficiencies in all cell lines (range 1.6–9.5%). Defined media increases efficiency to 23%. Interline variability may be seen. 25% beating by d8 increased to 70% by d16. Contractions observed till d70. Beating efficiency varied from 1% to 10% for various iPSC lines.
Kim et al.
cells is important for cardiac repair. Similarly, clinical trials performed using undifferentiated adult stem cells demonstrated their ability to help in functional recovery in patients, it is unclear if predifferentiating them to cardiogenic lineage could yield better outcomes. Adult Stem Cell-Derived Cardiomyocytes in Cardiac Repair Most of the clinical studies performed so far have used bone marrow-derived mononuclear cells (BMCs) for treating patients with acute or chronic infarcts [reviewed by Dimmeler et al. (24)]. In clinical setting >1,000 patients have been treated with bone marrow-derived cells—either unfractionated (most studies) or enriched in progenitor subpopulations—in numerous clinical trials worldwide (66). However efficacy [gauged by left ventricular ejection fraction (LVEF) or magnetic resonance imaging (MRI)] has been inconsistent and primary end points in the majority of large randomized controlled studies have not been met (66). Instead a modest but significant benefit was seen in meta-analysis of all published studies (1,64). A number of plausible explanations have been given for these modest effects, which include (a) impaired functional activity of BM-derived cells in heat failure patients, (b) engraftment and short-term homing in patients (18), (c) paracrine effects from transplanted cells, and (d) increased stiffness of the ischemic areas of the ventricular wall from increase cell mass (39).
300–400 μm microwells EBs generated maximum beating by d30.
5-Azacytidine enhanced efficiency, DMSO had no effect and low serum and BMP2 marginally improved efficiency. p38MAPK inhibitor enhanced beating efficiencies by twofolds. Beating 14 days postinduction.
Long-term culture enhanced CM differentiation.
(34,69) (56) (51)
Furthermore, in vivo cardiomyogenic differentiation abilities of these stem cells are extremely low, long-term beneficial effects are elusive (18,39). Indeed, Pozzobon et al. (85) failed to detect differentiation of CD133+ bone marrow stem cells (BMSCs) to cardiomyocytes either in vitro and in vivo. It could be suggested that milieu-directed differentiation may provide better answers to long-term beneficial effects. Following demonstration of such ability in cardiac resident stem cells in response to cardiac injury to form three major cardiovascular cell types, SCIPIO and CADUCEUS trials (70) were initiated to evaluate the concept of milieu-directed differentiation. In SCIPIO trial (clinical trials.gov: NCT00474461), patients with ische mic cardiomyopathy received an infusion of autologous cardiac stem cells (CSCs) isolated from atrial appendage, while LV dysfunction patients in CADUCEUS trial (clinical trials.gov: NCT0089336) received cardiospherederived cells (CDCs) (117). While the recently completed CADUCEUS trial did not improve LVEP (65a), initial results from 16 patients in SCIPIO trial (12) have shown improvements. In 14 CSC-treated patients, LVEF increased from 30.3% to 38.5% at 4 months, which did not change in control patients. Furthermore, in seven patients there was a significant reduction in infarct size by 24% at 4 months and 30% by 1 year. These initial results in patients are very encouraging, and suggesting that intracoronary infusion of autologous CSCs is effective in
6 mehta and shim
improving LV systolic function and reducing infarct size in patients although role of cardiomyogenesis by those CSCs remains to be determined. Despite a large pool of safety data and preclinical application of BMSCs in cardiac repair, their use in experimental infarction models has not been without problems (13,16). Injection of BMSCs into infarcted areas may differentiate into fibroblasts or form intramyocardial calcifications (61,127). Furthermore, efficiency of in vivo cardiac differentiation of BMSCs is debatable. To avoid such uncertainty, a number of groups have utilized differentiated adult stem cells in expecting better recovery in preclinical models. Our group transplanted differentiated CLCs into peri-infarct borders of infarcted rodent myocardium. In comparison to undifferentiated MSCs, CLCs prevented negative remodeling and improved hemo dynamic recovery (110). Furthermore, superiority of CLCs over MSCs was demonstrated in load-independent measurements of the end-systolic pressure–volume relationship and preload recruitable stroke work (100). Similarly, Behfar et al. (8) demonstrated that guided cells when delivered into infarcted myocardium of nude mice showed superior functional and structural benefits as compared to their unguided counterparts. They further demonstrated that transplantation of these cells significantly improved left ventricular fractional shortening (unguided BM vs. guided BM: -4.8 ± 2.1% vs. 5.2 ± 1.5%). Potapova et al. (84) demonstrated that cardiac differentiated hMSCs repaired a full thickness canine right ventricular defect as compared to unmanipulated hMSCs. In a recent study, Li et al. (61) compared the efficacy of induced BMSCs (iBMSCs) and uninduced BMSCs (uBMSCs). iBMSCs were cocultured with rat cardiomyocytes before being transplanted into border regions of rodent cardiac scar. Echocardiography results indicated near normal LVEF in iBMSC injected rats, but a lower LVEF in uBMSC groups at 4 weeks (iBMSC vs. uBMSC: 80% vs. 53%, p